⚠ New certification: Anthropic published its guide (version 1.0, November 2026) on 17 November 2026. The lessons and questions are a first draft built on that guide and will be refined.
CCDV-P
Claude Certified Developer (Professional)
Who it is for: Senior engineers who run Claude applications and agents in production: reliability, quality, cost, security, and fleets of coding agents.
70 questions · 120 minutes
25 lessons · 8 domains
Take a practice examHow do I take the exam?
- Your company must be part of the Claude Partner Network. Joining is free. If it is not yet, sign up for Kelsoft’s Anthropic Challenge and ask Cande.
- On the Anthropic Partner Academy, open the certification page, read the guide and the terms, and register. The list price is USD 175; checkout shows the final price, with any discount for your company's partner tier.
- On the Academy, choose Register and confirm your name matches your government-issued ID. You then go straight to Pearson, already signed in, to schedule and pay: online proctored or at a Pearson test centre. If you were already registered, do not register again: use “Go to my Pearson account” instead.
- You can cancel or reschedule up to 24 hours before. After that, the fee is forfeited.
- The credential lasts 12 months. Renewing on time is free: a short, non-proctored assessment on what has changed.
Mocks
Your preparation
Diagnostic
You haven't taken it yet. See where you stand in each domain.
Take the diagnostic Timed diagnostic (account required)Lessons
0 of 25 read
Next: Multi-agent patterns in production and how they fail at the seams
Drill
Series of 15 questions aimed at one domain.
Once you take the diagnostic, we'll suggest the domain that's best to practice.
Start a drillDomains
Advanced Agent Architecture and Orchestration 17%
- Multi-agent patterns in production and how they fail at the seams
- Durable execution, checkpoints, sessions and resumability
- Evolving between workflow and agent with data, and setting harness and framework standards
- Context and memory policies for long-running agents
Production Integration, Reliability, and Knowledge Grounding 20%
- Reliability, retries, circuit breakers and fallbacks that tell causes apart
- Scale and throughput, finding the binding constraint
- MCP in production and model or prompt changes without surprises
- Retrieval and grounding, diagnosing by stage and not making things up
Observability and Production Monitoring 12%
- What to instrument in a Claude system, and what not to store
- SLIs, SLOs, alerting and quality drift
- Trace analysis at scale, sampling, aggregation and failure cohorts
Advanced Debugging and Incident Response 10%
- Isolating where a failure originates, model, harness, tools or data
- Incidents, mitigate first and turn the learning into evaluations
Evaluation Engineering and Quality Assurance 12%
- Designing the evaluation system, golden sets, regression and release gates
- An LLM judge must be calibrated, rubrics, biases and agreement with people
- Measuring quality on live traffic, sampling, A/B tests and feedback
Cost and Latency Optimization at Scale 8%
- Cost architecture, model, route, cache and cap
- Latency engineering, first token or total, and what to move
Runtime Guardrails and Production Security 11%
- Runtime guardrails, layers, hooks and keeping the irreversible outside the model
- Agents facing untrusted content, the dangerous combination and defense in depth
- Agent identity, secrets without static keys and audit
Agent Fleet Operations for Software Delivery 10%
- Parallel agents, worktrees, decomposition and merge strategy
- When generation outpaces review, AI code review in shadow mode
- Agents with nobody watching, goals, schedules and event response
- Configuring coding agents across the whole organization